Pilot Design Based On Compressed Sensing In PLC System

The complete infrastructure of grid is used for data communication in the power line communication(PLC)technology,which reduces the expenditure on capital construction funds of the communication network and improves the resource utilization,so it has great development prospects.The channel estimation is one of the key technologies in PLC systems,and the merits of channel estimation are closely related to pilot design.Because of the sparse nature of PLC channel,the compressed sensing(CS)technology can be used to design the pilot of sparse channel,which saves the spectrum resource greatly.However,the complex attenuation characteristics and noise interference of power line channels have seriously affected the quality of data communication.Therefore,considering the impulse noise in the power line,it is necessary to research the CS reconstruction algorithm of the anti-impulse interference.And the main research in this thesis is as follows:1.The CS pilot design algorithms are analyzed,which include the Stochastic Search Schemes(SSS)and Fast Greedy Search(FGS)algorithm.Due to their high computational complexity,a two-dimensional Search Binary Artificial Bee Colony(TS-BABC)algorithm is proposed,which maps the pilot design process to the binary discrete domain and performs the pilot position replacement through a two-dimensional neighborhood cross search.Compared with the SSS and FGS algorithm,the TS-BABC algorithm is less affected by the change of the number of pilots,and it has lower complexity and faster convergence speed without losing the pilot performance.2.The reconstruction algorithms of anti-impulse noise based on the minimum Lorentzian norm are analyzed,which can obtain reliable performance of signal reconstruction in the impulse noise environments.However,the negative gradient of the residual is used to update the signal in these algorithms,which introduces the effect of noise in the estimate of signal,so the Lorentzian Double Threshold Residual Pursuit(LDTRP)algorithm is proposed.The algorithm corrects the estimate of signal by estimating the reconstructed error in the iteration and uses the residual to set the outliers threshold.The existing algorithms are simulated and compared under the impulse noise environment,the LDTRP algorithm can obtain a remarkable improvement of performance in three aspects: the change of impulse noise numbers,impulse noise power,and the sparsity of signal.3.The OFDM-PLC system link is established,the TS-BABC algorithm is used for pilot design at the sender,and the LDTRP algorithm is used to restore the power line channel at the receiver.Experimental results show that the proposed scheme can obtain considerable BER performance in both gaussian noise and impulsive noise environments.